BaseSpaceR Adrian Alexa [email protected]

© 2011 Illumina, Inc. All rights reserved.Illumina, illuminaDx, BaseSpace, BeadArray, BeadXpress, cBot, CSPro, DASL, DesignStudio, Eco, GAIIx, Genetic Energy, Genome Analyzer, GenomeStudio, GoldenGate, HiScan, HiSeq, Infinium, iSelect, MiSeq, Nextera, Sentrix, SeqMonitor, Solexa, TruSeq, VeraCode, the pumpkin orange color, and the Genetic Energy streaming bases design are trademarks or registered trademarks of Illumina, Inc. All other brands and names contained herein are the property of their respective owners. 

BaseSpaceRAdrian [email protected]

2

BaseSpace - Plug and Play Genomic Cloud SolutionFrom Sample to Biological Insight

1. Seamless Instrument Integration

2. Harness the Internet

3. Accessible to EveryoneData transferred as the instrument runs

Data available in BaseSpace within minutes of a run finishing

Automatic de-multiplexing and FASTQ generation

3

BaseSpaceData storage, analysis and collaboration.

Almost 40,000 Instrument Runs Streamed to BaseSpace by April 2013

4

UIDev

PortalApp App

Users

Developers

SaaS

Instruments

PaaS

OpenStack

IaaS

Secondary Analysis

Visualization

Biological InferenceValidation

Methods Development

App App App App

BaseSpace API

BaseSpace APIREST API and Native Apps

5

BaseSpace APIData Model

6

BaseSpaceR = BaseSpace + R + BioconductorA translation layer between BaseSpace’ REST API and R data structures

Cloud based content management system, facilitating the storage and sharing of genomic data.

Rich environment of statistical and data analysis tools for high-throughput genomic data.

Persistent connection with the REST server and support for the REST API query parameters.

Vectorized operations in line with the R semantic. Allows for queries across multiple Projects, Samples, AppResults, Files, etc.

S4 class system used to represent the BaseSpace data model.

Integration with Bioconductor libraries and data containers [working on it…].

Portability on most platforms: Linux, Windows and Mac OS X.Fe

atu

res

7

BaseSpaceRREST API and R data structures

REST API R APIGET https://api.basespace.illumina.com/v1pre3/samples/16018/files?Extensions=gz&Offset=3&Limit=1

8

Q-score distributionAccess to the FASTQ files

1) Authentication [3 lines of code]– User needs to interact with the BaseSpace UI (or via a web server).

> aAuth <- AppAuth(client_id = "5b123858536b473ba740e9a9eb0abf64",+ client_secret = "b3168bf65bf543f3b6e7f087856922df",+ scope = "CREATE GLOBAL, BROWSE GLOBAL, CREATE PROJECTS")

Launching browser for OAuth authentication... > requestAccessToken(aAuth)

Access token successfully acquired!

> aAuthObject of class "AppAuth" with:

9

Q-score distributionAccess to the FASTQ files

> myProj <- listProjects(aAuth)> data.frame(Name = Name(myProj), Id = Id(myProj)) Name Id 1 BaseSpaceDemo 2 2 Cancer Sequencing Demo 4 3 HiSeq 2500 7 ...............................> sampl <- listSamples(aAuth, projectId = 2, Limit = 1)> inSample <- Samples(sampl, simplify = TRUE)> inSample#Samples object:

1) Authentication [3 lines of code]– User needs to interact with the BaseSpace UI (or via a web server).

2) Select a sample (collection of FASTQ files) from the Project of your choice [3 lines]

10

Q-score distributionAccess to the FASTQ files

> f <- listFiles(inSample, Extensions = ".gz")> idx <- grep("_R(1|2)_", Name(f))> outDir <- paste("Sample", Id(inSample), sep = "_")> getFiles(aAuth, id = Id(f)[idx], destDir = outDir, verbose = TRUE)Downloading 4 files in directory: Sample_16018

Downloading file: data/intensities/basecalls/s_G1_L001_R1_001.fastq.1.gz........................................................................

> file.exists(file.path(outDir, f$Path[idx]))[1] TRUE TRUE TRUE TRUE

1) Authentication [3 lines of code]– User needs to interact with the BaseSpace UI (or via a web server).

2) Select a sample (collection of FASTQ files) from the Project of your choice [3 lines]

3) Download the files (FASTQs in our case) [4 lines]

11

Q-score distributionAccess to the FASTQ files

1) Authentication [3 lines of code]– User needs to interact with the BaseSpace UI (or via a web server).

2) Select a sample (collection of FASTQ files) from the Project of your choice [3 lines]

3) Download the files (FASTQs in our case) [4 lines]

4) Process the downloaded files and compute the stats […]

> library(ShortRead)> source("QscoreApp-functions.R")> qtab <- lapply(floc, getQscoreCounts)> idxR1 <- grep("_R1_", names(floc), fixed = TRUE)> idxR2 <- grep("_R2_", names(floc), fixed = TRUE)> x <- getQscoreStats(cbind(Reduce("+", qtab[idxR1]), Reduce("+", qtab[idxR2])))> ylim <- range(x) + c(-2L, 2L)> plot(x = seq_len(nrow(x)), type = "n", ylim = ylim,+ xlab = "Cycle", ylab = "Q-score",+ main = "Q-scores statistics")> sx <- apply(x[, c("5%", "95%")], 2, function(x) smooth.spline(x)$y)> sx[, "95%"] <- pmax(sx[, "95%"], x[, "median"])> polygon(c(1L:nrow(x), nrow(x):1L), c(sx[, "95%"], rev(sx[, "5%"])), col = "#CCEBC580", border = NA)> matpoints(sx, type = "l", lwd = .5, lty = 2, col = "black")> lines(x[, "mean"], lwd = 2, col = "red")> lines(x[, "median"], lwd = 2, col = "black")

12

Q-score distributionAccess to the FASTQ files

1) Authentication [3 lines of code]– User needs to interact with the BaseSpace UI (or via a web server).

2) Select a sample (collection of FASTQ files) from the Project of your choice [3 lines]

3) Download the files (FASTQs in our case) [4 lines]

4) Process the downloaded files and compute the stats […]5) Upload results back to BaseSpace [~10 lines]

– Results are collection of files for now, minimal visualisation.

13

Amplification of distal chr8q.

Homozygous deletion of part of chr8p.

Location of centromeres indicated by vertical dotted lines

Detect amplifications and deletions in cancer samples

Data can be obtained from a single MiSeq runs (one for tumor and one for normal or even both on a flowcel).

Typical analysis requires only the coverage data and this can be directly obtained using a REST method.

Corrects for tumor ploidy and purity.

Copy Number AbnormalitiesAccessing the coverage via a high-level REST method

14

RConsoleExploring BaseSpace data using RStudio

15

What’s next

Facilitate the use of Bioconductor packages – there is much to gain if as many Bioconductor packages as possible can consume data (directly) from BaseSpace.

Introduce high-level methods (REST or R API) for random access to BAMs, VCFs, metric data, etc. One can already use Rsamtools for indexed BAMs.

R level methods to facilitate RNAseq, ChipSeq, etc. analyses.

BaseSpace Data Central – publicly available data – most of it will be data coming from our latest instruments, chemistry, workflows.

16

Resources

Tutorials, videos, whitepapers and other educational material:http://www.illumina.com/software/basespace/basespace-education.ilmn

BaseSpace homepage:https://basespace.illumina.com

BaseSpace developer portal:https://developer.basespace.illumina.com

Bio-IT World Asia presentation:https://dl.dropboxusercontent.com/u/14162259/BioITAsia_MJJ.pptx

More documentation on BaseSpace:http://support.illumina.com/sequencing/sequencing_software/basespace/documentation.ilmn

BaseSpaceR homepage on Bioconductor:http://bioconductor.org/packages/release/bioc/html/BaseSpaceR.html